INDEX

Explanations

patterns with specific characters and symbols, potentially related to coding or data matching

oai_token-act-pair · gpt-3.5-turbo

specific patterns related to regular expressions or special characters

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Showtime

-0.82

 Samar

-0.80

 methamphetamine

-0.78

 Snapchat

-0.71

 facilit

-0.66

Tel

-0.64

 predictably

-0.64

 executive

-0.63

Pel

-0.61

 DACA

-0.61

POSITIVE LOGITS

3.86

^^

2.69

2.29

^{

2.17

^^^^

2.01

Â«

1.45

**

1.40

\-

1.39

 âĸł

1.34

1.33

Activations Density 0.027%

patterns with specific characters and symbols, potentially related to coding or data matching

specific patterns related to regular expressions or special characters

No Comments

No Known Activations